anaerobic infections of the abdomen and pelvis

Anaerobic infections of the abdomen and pelvis

Sydney M. Finegold, M.D.*

*Department of Medicine, Wads-worth Veterans Administration Hos-pital and the University of Califor-nia Medical Center, Los Angeles, California.

Recently, there has been an increased aware-ness of the role of anaerobes in clinical infections. The purpose of this paper is to emphasize the fre-quency with which anaerobes are involved in in-fections of the abdomen and pelvis, and to illus-trate some of the clinical problems.

First, I would like to present data from an ex-cellent study by Gorbach et al.1 Forty-six cases were studied; predisposing conditions included abdominal trauma (22 patients), carcinoma (7 pa-tients), postoperative infections (7 patients), per-forated appendix (4 patients), hepatic cirrhosis (3 patients), and peritoneal dialysis (3 patients). Thirty-two patients had intraabdominal abscesses, 10 had generalized peritonitis, and four had mis-cellaneous conditions (retroperitoneal and pro-static abscesses, infected aortic graft). Of 43 speci-mens of purulent abdominal material studied, 33 yielded both anaerobes and aerobes, seven grew anaerobes alone and in three, only aerobes were isolated. There was an average of five organisms per specimen (range 1 to 13), with a ratio of three anaerobes to two aerobes. Each organism had to be obtained in pure culture in order to be ac-curately identified, a time-consuming process with anaerobes, which often grow slowly.

The specific organisms are listed in Table 1. Gram-negative nonsporeforming anaerobic rods predominated, reflecting the fact that Bacteroides

113

require permission. on May 28, 2022. For personal use only. All other useswww.ccjm.orgDownloaded from

http://www.ccjm.org/

114 Cleveland Clinic Quarterly Vol. 42, No. 1

Table 1. Anaerobic microorganisms isolated from 43 intra-

abdominal wounds

Gram-negat ive nonsporing rods Bacteroides 36

B. fragilis 28 B. melaninogenicus 4 B. clostridiiformis 2 B. biacutus 1 Bacteroides species 1

Fusobacter ium 6 F. varium 1 F. necrophorum 1 Fusobacterium species 4

Cocc i Peptostreptococci 11 Peptococcus 3

P. magnus 2 P. prevotii 1

Vei l lone l la 1 Gram-pos i t ive nonsporing rods

Eubacter ium 10 E. lentum 6 E. filamentosum 2 E. multiforme 1 E. limosum 1

Propionibacter ium 2 Lactobacillus catenaforme 2 Bif idobacterium 1 Clostridia 31

C. ramosum 3 C. perfringens 3 C. butyricum 2 C. bifermentans 2 C. sporogenes 2 C. innocuum 2 C. sordellii 1 C. fallax 1 C. beijerinckii 1 C. barati 1 C. difficile 1 C. ghoni 1 C. sphenoides 1 Clostridium species 10

fragilis is the most prevalent member of the normal flora of the bowel. This organism was isolated 28 times along with small numbers of other gram-neg-ative nonsporeforming anaerobes of the genus Bacteroides and of the genus

Fusobacterium. Of special significance is the fact that B. fragilis is much more resistant to antimicrobial agents than other anaerobic bacteria. Fusobacte-rium varium, isolated only one time in this study, may also be resistant, al-though not typically to the extent that B. fragilis is. However, one third of strains of F. varium that we have stud-ied have been resistant to clindamycin, an otherwise excellent agent for anaer-obic infections. There were a smaller number of anaerobic cocci of various types and of gram-positive nonspore-forming anaerobic rods. Some mem-bers of the genus Eubacterium may also be resistant to clindamycin. There were a large number of Clostridia, a total of 31 isolates. Clostridium ramo-sum, relatively commonly involved in anaerobic infections of all types, was seen here as frequently as Clostridium perfringens. C. ramosum is the third member of the anaerobic group which tends to be resistant to antimicrobial agents including, to some extent, clin-damycin and even penicillin. Ten per-cent to 15% of strains of these other Clostridia species may also be resistant to clindamycin as well. T h e aerobic and facultative forms isolated in this series are shown in Table 2. Esche-

Table 2. Aerobic and facultative micro-organisms isolated from 46 intra-

abdominal infections

Escherichia coli 28 Staphylococcus epidermidis 15 Proteus 10 Klebs ie l la 9 Pseudomonas 8 Enterobacter 8 Cand ida 4 Streptococcus 3 Enterococcus 2 Staphylococcus aureus 1



Spring 1975 Anaerobic infections 115

richia coli predominated as expected. A number of patients had organisms not commonly found in significant numbers among the normal bowel flora such as Klebsiella, Pseudomonas, and Staphylococcus aureus, reflecting, most likely, some prior antimicrobial ther-apy or hospital experience or both.

There were 13 instances in which bacteremia was present. In 11 of these, anaerobes were isolated in pure cul-ture, and in the other two there was a mixture of anaerobes and aerobes. The specific organisms responsible for bac-teremia in this series of intraabdomi-nal infections are shown in Table 3. Again, B. fragilis predominated. Also present were Bacteroides melaninogen-icus, several Clostridia, a fusiform, and Klebsiella isolated in two cases in mixed culture along with anaerobes. Two C l o s t r i d i a were resistant to clin-damycin, one of them C. perfringens, although C. perfringens is usually sus-ceptible.

Other categories of intraabdominal infection

Figure 1 depicts a subdiaphragmatic abscess shown by a combined liver-lung scan; there is an abnormally large space between the liver and the lung. Most of these abscesses are secondary to pathology in the bowel, and there-fore anaerobes are very commonly in-volved in subphrenic abscess. The com-plex bacteriology that may be encoun-tered in this type of process and in in-traabdominal infections in general is illustrated in Table 4. Four aerobic or-ganisms were isolated (the diphthe-roids probably representing a contami-nant) together with a yeast and then a total of nine different anaerobic orga-nisms. This type of specimen is a large project for the clinical laboratory.

Table 3. Microorganisms in 13 cases of septicemia associated with intra-

abdominal infections*

Bacteroides fragilis 8 B. melaninogenicus 1 Clostridium perfringens 2 C. tertium 1 C. difficile 1 Fusobacterium nucleatum 1 Klebsie l la 2

* T w o patients had B. fragilis and Klebsie l la ; one patient h a d 2 strains of Clostridia.

Treatment must be based on clinical acumen, with the information derived from a direct Gram stain, and with recognition that it will take time to get definitive bacteriologic results. Figure 2 shows an appendiceal abscess indenting the cecum outlined by a barium enema. Figure 3 represents a case of diverticulitis as shown by barium enema. Although some diverti-cula are filled, one had previously ruptured and barium extravasated into a large abscess adjacent to the colon. This process, as well as appen-diceal infections, involves anaerobic bacteria essentially 100% of the time. Failure to recover anaerobes from this type of infection should prompt care-ful réévaluation of techniques. The laboratory may be at fault, but, more often, the specimen has not been transported properly (exposure to air during transport may kill anaerobes).

Figure 4 is an example of postopera-tive bacterial synergistic gangrene which occurs after various types of sur-gery, but is most common after abdo-minal surgery (particularly bowel sur-gery) especially when retention sutures are required, as in this case. The clini-cal picture is typical. Severe pain is as-sociated with this process. In the classic case, there are three zones: a central ne-




Fig. 1. Combined l iver- lung scan showing an a b n o r m a l l y large area of decreased pe r fus ion between the l iver a n d the lung . T h i s proved to be a subphren i c abscess.

Table 4. Subphrenic abscess; organisms isolated in one case*

A e r o b e s Escherichia coli Klebsiella species Proteus mirabilis D i p h t h e r o i d s

A n a e r o b e s Bacteroides fragilis P e p t o s t r e p t o c o c c u s P e p t o c o c c u s Eubacterium multiforme E. lentum Clostridium bifermentans C. barati C. innocuum V e i l l o n e l l a

F u n g i Candida albicans

* F r o m case wi th lung-liver scan shown in Figure 1.

Fig. 2. B a r i u m enema showing extr insic com-pression of t h e cecum secondary to an appen -diceal abscess.




Fig. 3. Diver t icul i t is of the sigmoid colon wi th pe r fo ra t ion . T h e small a r rows po in t to diver-t icula. T h e large arrow poin t s to b a r i u m ex-t ravasat ing t h r o u g h a pe r fo ra t ed d iver t i cu lum in to a large per icolonic abscess.

crotic zone surrounded by a purplish zone of discoloration, and outside of that a lighter red zone of inflamma-tion. If material is cultured from the central necrotic area, even where there is frank pus, recovery of S. aureus only is likely. However, if the active margin of the inflammatory area is cultured, microaerophilic streptococci in addi-tion to S. aureus would be recovered. Those are the classic organisms that are recovered. On occasion, instead of 5. aureus, a gram-negative rod such as Proteus or Klebsiella may be recovered, and in lieu of a microaerophilic strep-tococcus, an anaerobic streptococcus. This is a mixed infection and the synergism is significant. If this patient were treated with methicillin alone, on the assumption that the staphylococcus isolated from the center of the necrotic

lesion was the true infecting organism, chances are that the patient would not respond or would respond in subop-timal fashion. However, if therapy was extended to also cover the microaero-philic streptococcus (often resistant to an agent such as methicillin), the re-sponse would be good provided surgi-cal drainage and excision of necrotic tissue were also carried out. These are virulent infections and too often go on, as this one did, to complete wound de-hiscence (Fig. 5).2

Figure 6 depicts a lateral liver scan with a posterior defect due to a liver abscess in a 13-year-old girl who had a rout ine appendectomy 1 month earlier. One week postoperatively, low grade fever developed which persisted and became more pronounced. She had anorexia, lost weight and was finally readmitted to the hospital. Blood cul-tures yielded B. fragilis on two occa-

Fig. 4. P h o t o g r a p h of postopera t ive abdomi -nal w o u n d showing edema of the edges of the w o u n d a n d gangrenous changcs of the skin a r o u n d stainless steel sutures. An i r r igat ion ca the te r is in place. (Reproduced wi th permis-sion of a u t h o r , d e j o n g h DS, Smith JP , T h o m a GW: Postopera t ive synergistic gangrene . J A M A 200: 557-559, 1967.)



118 Cleveland Clinic Quarterly

Fig. 5. L a t e r p h o t o g r a p h of a b d o m i n a l w o u n d showing dehiscence, fou r en te rocu taneous fis-tu lae , and a large vesicocutaneous fistula. (Rep roduced wi th permiss ion of a u t h o r , de-J o n g h DS, Smith JP , T h o m a G W : Postopera-tive synergistic gangrene . J A M A 200: 557-559, 1967.)

Fig. 6. Latera l liver scan showing a r a t h e r cir-cumscr ibed , r o u n d e d area of decreased radio-Iucency in the pos ter ior - infer ior aspect of the r igh t lobe of the liver. T h i s was proved to be caused by a liver abscess.

sions. Her anteroposterior liver scan was entirely normal. T h e abscess was drained and B. fragilis was recovered in pure culture from the pus.

Vol. 42, No. 1

Our experiences with liver abscess at the Wadsworth Veterans Hospital and U.C.L.A. over a 10-year period include a total of 47 cases.3 In 19 cases, only aerobes were isolated, in six there were mixtures of aerobes and anaerobes, and in 15, only anaerobes were isolated (Table 5). Anaerobes were isolated from 21 of 45 cases from which data were available. If one subtracts the five cases which yielded no growth, at least two of which showed organisms on di-rect smear (and therefore were presum-ably anaerobic), this would mean that over half of our cases of pyogenic liver abscess involved anaerobes. Among the 19 cases which showed only aerobes, several grew only small numbers of or-ganisms such as Staphylococcus epider-midis, Micrococcus or diphtheroids which undoubtedly represented con-tamination. This was a retrospective study and only a portion of the cul-tures were performed in our own lab-oratory. It is not known if all speci-mens actually were submitted for an-aerobic culture or how effectively those that were submitted for anaerobic cul-ture were studied. There was a signifi-cant problem with regard to recovering anaerobes in our clinical laboratory for

Table 5. Pyogenic liver abscess (Wads-worth Veterans Hospital,

U.C.L.A., 1960-1970)

No. of cases

A e r o b e s

M i x e d a e r o b e s a n d a n a e r -obes

A n a e r o b e s

N o g r o w t h 5 N o c u l t u r e t a k e n 1 C h a r t n o t l o c a t e d 1

T o t a l 47 (4)

H)

6 (2)

15 (2)




a period of time. If a prospective study were done with proper culture of all specimens, it is likely that at least two thirds of pyogenic liver abscesses would be shown to be due to anaerobic or-ganisms.

In addition to the 21 cases of an-aerobic pyogenic liver abscess noted above, there were four other cases from other hospitals that we had an opportunity to see and culture. Thus, a total of 25 cases of anaerobic liver abscess were analyzed. T h e clin-ical picture was not really different from pyogenic liver abscess in gen-eral. T h e predisposing conditions, such as underlying bowel or gall-bladder disease, were what one ex-pects to see. T h e general and local-izing symptomatology was not unique; neither were the physical nor labora-tory findings. This coincides with the idea that pyogenic liver abscess is frequently an anaerobic process, and one should not expect a different clinical or laboratory picture from that described in the past for liver abscess, which has been thought to be mostly due to aerobes. General bacteriologic data for these 25 cases are noted in Table 6. Twelve cases yielded a single anaerobe while 13 had anaerobes in mixed culture. Five of the latter had only anaerobes (two or more), but no aerobes or facultatives; the other eight had mixtures of an-aerobes and aerobes. T h e specific an-aerobes recovered were predomi-nantly cocci, either anaerobic cocci, streptococci, or microaerophilic strep-tococci. B. fragilis, the resistant form that has been discussed, was seen with a certain frequency. There were also other gram-negative anaerobic rods, including Fusobacterium nucleatum and one Actinomyces.

Table 6. Anaerobic pyogenic liver abscess; 25 cases (1960-1970)

Bacteriological data, general Patients

S ingle anaerobe 12 M i x e d cu l ture 13

Mixed—anaerobes only 5 Mixed—anaerobes and aerobes 8

Table 7 is taken from a 1927 report4

and deals with puerperal infection of various types (e.g., endometritis, pelvic cellulitis, peritonitis, associated bac-teremia). T h e third column shows the number of aerobic bacteria iso-lated f rom these various types of puerperal infection and column 4, the number of anaerobic streptococci. One can readily see that anaerobic streptococci outnumber the aerobic organisms in this series. Wi th mod-ern techniques we would expect to find, in addition, a certain number of anaerobic gram-negative bacilli and of Clostridia. Table 8 is a list of vari-ous female genital tract infections that may involve anaerobic bacteria. Anaerobes are very commonly in-volved, not only in puerperal infec-tion, but also in postabortal sepsis, in postoperative obstetric and gyneco-logic infections, and in a variety of other infections. T h e specific orga-nisms involved are listed in Table 9. In contrast to the intraabdominal in-fections in which gram-negative anaer-obic bacilli were the dominant orga-nisms and B. fragilis stood out particu-larly; in female genital tract infections the anaerobic cocci are much more common. This reflects the fact that in this site the anaerobic cocci are much more commonly found as indigenous flora, whereas in the bowel, the Bacte-roides group, and B. fragilis in particu-lar, is dominant. In female genital tract




Table 7. Types of puerperal infection*

No. of cases of No. of No. of admissions 7-1-24 to 9-1-26 No. of deliveries puerperal infection deaths

2194 1913 45 10

Type of infection No. of Aerobic cases bacteria

Anaerobic strepto-

cocci Mixed

bacteria Negative cultures

Mor-tality

A c u t e endometri t i s 42 11 27 3 1 0 Pe lv ic cel lul it is 11 4 6 1 0 0 Peritonit is 7 1 4 2 0 6 Pe lv ic abscess 3 1 1 1 0 0 Pe lv ic thrombophlebi t i s 9 0 6 0 3 3 Bacteremia 15 4 10 1 0 1

* D a t a obta ined from Schwarz O H , D i e c k m a n n W J : Puerperal infect ion due to anaerobic s treptococci . A m J Obs te t Gyneco l 13: 467-485, 1927.

Table 8. Anaerobic infections of the female genital tract

Endometr i t i s , pyometra Myometr i t i s Parametr i t i s Pe lv ic cel lul i t is , abscess Pelv ic thrombophlebi t i s Peritonit is , intraabdominal abscess Bacteremia , w i t h or w i t h o u t shock or intra-

vascular hemolysis M e t a s t a t i c infect ion—particularly of lung,

l iver, brain, heart valves, k idney W o u n d infect ion or abscess V u l v a r abscess Barthol init is , Barthol in abscess Skenit is , Skene's abscess Paracl i toroidal abscess Vagin i t i s , abscess of vaginal wal l , para-

vag ina l abscess Salpingit i s , tubal abscess O v a r i a n or tubo-ovarian abscess Abscesses of adjacent parts—groin, peri-

rectal or ischiorectal areas, paraurethral , leg , abdomina l wal l

Chor ioamnioni t i s Fetal e m p h y s e m a

Intrauter ine or neonata l pneumonia , sepsis

infections, one encounters Peptostrep-tococcus, Peptococcus, microaerophilic cocci and some gram-negative cocci more commonly than other organisms.

However, B. fragilis and other gram-negative anaerobic rods are seen as well. Fortunately, C. perfringens is not commonly encountered as it may pro-duce devastating infection. Occasion-

Table 9. Organisms involved in female genital tract infections

Anaerobes Peptostreptococcus Peptococcus Microaerophi l i c gram-pos i t ive cocci Ve i l l one l la Bacteroides fragilis Bacteroides melaninogenicus O t h e r Bacteroides species Fusobacterium necrophorum O t h e r Fusobacter ium species Clostridium perfringens Actinomyces israelii

O t h e r organisms Escherichia coli O t h e r Enterobacter iaceae Group A beta -hemolyt i c s treptococci Enterococc i O t h e r streptococci Staphylococcus aureus G o n o c o c c u s

Haemophilus (Corynebacterium) vaginalis Proteus Pseudomonas




ally, other Clostridia and nonspore-fo rming gram-positive anaerobes may be recovered. Among the aerobic organisms, E. coli and g roup A strep-tococci dominate , bu t others may be seen in addi t ion .

As stated before, anaerobic bacteria prevalent as indigenous flora on muco-sal surfaces become involved in infec-tions of the bowel and the female geni-tal tract. T h i s is t rue all th rough the body and the number of these orga-nisms present as normal flora is qu i te impressive. I n the bowel, anaerobes o u t n u m b e r aerobic or facultat ive forms a thousand to one. E. coli, which we commonly th ink of in terms of bowel organisms, constitutes only one t en th of 1% of the total bowel flora. T h e same th ing is t rue to a lesser ex-tent in other areas. I n the female geni-tal tract, anaerobes o u t n u m b e r faculta-tive or aerobic forms ten to one; this is a common ra t io in other sites in the body. So, if one is to obta in a specimen to document the bacteriology of an in-fectious process, part icular ly an anaer-obic infect ion, one must avoid the

normal flora. I n the case of intraab-domina l infection, this is usually not too difficult a p rob lem unless there is an open communica t ion between the bowel and the per i toneal cavity or abscess. I n the female genital tract, it may be much more of a problem. An interest ing approach to this p rob lem was repor ted by Gui lbeau and Schaub5

(Fig. 7). It is impor t an t to posit ion the pa t ient carefully and get good expo-sure by proper use of a vaginal specu-lum. T h e vagina and cervical os are de-contamina ted with an iodophor . T h e authors used a sterile steel ins t rument over which was stretched a sterile rub-ber finger cot tha t was tied in place. T h e ins t rument was inserted in to the endocervical canal; then a stylet was pushed forward to perfora te the finger cot which would then retract. O n e could then utilize plastic t ub ing at-tached to a syringe or a swab to ob ta in contents f rom the endometr ia l cavity.

Ano the r impor t an t consideration for the clinician, in order to obta in relia-ble bacteriologic results in anaerobic infections, is to t ranspor t the specimen

Fig. 7. T e c h n i q u e for ob t a in ing endomet r i a l c u l t u r e relat ively f ree of n o r m a l flora. See text for descr ipt ion. (From Gu i lbeau JA J r , Schaub IG: Ute r ine cu l tu re technique . A s imple m e t h o d for avoiding c o n t a m i n a t i o n by cervical and vaginal flora. Am J Obs te t Gynecol 58: 407-410, 1949.)




to the laboratory in such a way as to avoid loss of organisms. When anaer-obes first leave the body in various dis-charges, they are quite sensitive to air exposure. It is therefore extremely im-portant that the specimen be placed under anaerobic conditions as promptly as possible. One useful tech-nique involves the use of a tube which has been gassed out with oxygen-free gas and stoppered with a butyl rubber stopper, butyl rubber being imper-vious to gas. T h e specimen is obtained by syringe, then all air is expelled from the syringe and needle and the con-tents of the syringe are discharged into the gassed-out tube. Another simpler and much less expensive approach is to transport the specimen directly to the laboratory in the syringe. T h e needle is capped with a rubber stopper. Speci-mens in syringes must be set up in an-aerobic culture within 20 to 30 min-utes. If it is absolutely necessary to use a swab, which it is on occasion, then it is best to use a swab which has been put u p in an anaerobic atmosphere. T h e swab should be made as wet as possible with the specimen and then put into a tube of semi-solid Carey and Blair transport medium, also in an anaerobic atmosphere.

T h e two major aspects of therapy for these infections are medical and surgical. T h e latter is the more impor-tant of the two. Drainage of collections of pus and excision of necrotic tissue, both of which are very commonly en-countered in anaerobic infections, are crucial to the patient's survival. T h e second aspect is antimicrobial therapy. There are five drugs or groups of drugs which are useful. Penicillin is active against most anaerobes with the ex-ception of B. fragilis. However, B. fra-

gilis is the anaerobe most commonly encountered in clinical infections. Car-benicillin, by virtue of the fact that high blood levels may be achieved safely, may encompass in its spectrum most strains of B. fragilis. Tetracy-clines are no longer satisfactory since a significant percentage of anaerobes of all types are now resistant. Doxycy-cline and minocycline are more active, but there are forms resistant to those agents. Chloramphenicol is active against all anaerobes and I would re-gard it as the single most dependable agent for treatment of anaerobic infec-tions. Chloramphenicol is the drug of choice in a critically ill patient with an anaerobic infection wherein the bacteriologic data are not yet avail-able. Clindamycin is an excellent drug, active against most anaerobes. One third of strains of F. varium, an un-common pathogen, are resistant to clindamycin, as are 10% to 15% of strains of a number of species of Clos-tridium other than C. perfringens. Therefore, in serious intraabdominal and female genital infections, where Clostridia are relatively commonly en-countered, penicillin would be needed to supplement clindamycin to cover that group of organisms, or one should use another agent such as chloram-phenicol. Metronidazole (Flagyl) is currently an experimental agent in anaerobic infection, but looks quite good in vitro. Preliminary clinical ex-periences are encouraging. It has the advantage of being remarkably bac-tericidal, much more so then even the penicillins.

Summary and conclusions

Proper collection and culture tech-niques reveal that anaerobes are the




most common pathogens associated with intraabdominal and pelvic infec-tions, including peritonitis, subphrenic and other loculated abscesses, hepatic abscess, and most puerperal and gyn-ecologic infections. Presumptive anti-biotic therapy, based on clinical aware-ness and the Gram-stain, should in-clude coverage for these anaerobes, as well as for the more easily recognized aerobic pathogens often present. Surgi-cal therapy, including drainage of ab-scesses and excision of necrotic tissue, in addition to selection of proper anti-biotics, is usually the key to successful therapy.

References 1. Gorbach SL, Thadepal l i H, Norsen J: Anae-

robic microorganisms in intraabdominal in-fections, in Anaerobic Bacteria: Role in Disease. Balows A, DeHaan RM, Dowell VR Jr, Guze LB, eds. Springfield, Illinois, Charles C Thomas, 1974, p p 399-407.

2. dejongh DS, Smith JP, T h o m a GW: Post-operative synergistic gangrene. JAMA 200: 557-559,1967.

3. Sabbaj J, Sutter VL, Finegold SM: Anaero-bic pyogenic liver abscess. A n n Intern Med 77: 627-638, 1972.

4. Schwarz OH, Dieckmann WJ: Puerperal in-fection due to anaerobic streptococci. Am J Obstet Gynecol 13: 467-485, 1927.

5. Guilbeau JA Jr, Schaub IG: Uterine culture technique. A simple method for avoiding contamination by cervical and vaginal flora. Am J Obstet Gynecol 58: 407-410, 1949.



anaerobic infections of the abdomen and pelvis

Documents